Hydrol. Earth Syst. Sci. Discuss., 3, 13151338, 2006 www.hydrol-earth-syst-sci-discuss.net/3/1315/2006/ Author(s) 2006.

. This work is licensed under a Creative Commons License.

Hydrology and Earth System Sciences Discussions

Papers published in Hydrology and Earth System Sciences Discussions are under open-access review for the journal Hydrology and Earth System Sciences

Klinkenberg eect for gas permeability and its comparison to water permeability for porous sedimentary rocks
W. Tanikawa1 and T. Shimamoto2
1

Kochi Institute for Core Sample Research, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), 200 Monobe-otsu, Kochi, Japan. 2 Department of Geology and Mineralogy, Faculty of Science, Kyoto University, Kitashirakawa-Oiwake-Cho, Sakyo-ku, Kyoto, Japan. Received: 20 April 2006 Accepted: 30 May 2006 Published: 7 July 2006 Correspondence to: W. Tanikawa (tanikawa@jamstec.go.jp)

1315

Abstract The dierence between gas and water permeabilities is signicant not only for solving gas-water two-phase ow problems, but also for quick measurements of permeability using gas as pore uid. We have measured intrinsic permeability of sedimentary rocks from the Western Foothills of Taiwan, using nitrogen gas and distilled water as pore uids, during several eective-pressure cycling tests at room temperature. The observed dierence in gas and water permeabilities has been analyzed in view of the Klinkenberg eect. This eect is due to slip ow of gas at pore walls which enhances gas ow when pore sizes are very small. Experimental results show (1) that gas permeability is larger than water permeability by several times to one order of magnitude, (2) that gas permeability increases with increasing pore pressure, and (3) that water permeability slightly increases with increasing pore-pressure gradient across the specimen. The results (1) and (2) can be explained by Klinkenberg eect quantitatively with an empirical power law for Klinkenberg constant. Thus water permeability can be estimated from gas permeability. The Klinkenberg eect is important when permeability is lower than 1018 m2 and at low dierential pore pressures, and its correction is essential for estimating water permeability from the measurement of gas permeability. A simple Bingham-ow model of pore water can explain the overall trend of the result (3) above. More sophisticated models with a pore-size distribution and with realistic rheology of water lm is needed to account for the observed deviation from Darcys law. 1 Introduction

10

15

20

25

Permeability (or intrinsic permeability) is an important parameter to control the uid ow systems at depths. Permeability would be wide ranges from 1012 to less than 1023 m2 for the various rock type and depth condition (i.e. Neuzil, 1994; Wang, 2000). Several sedimentary rocks indicate strong anisotropy in permeability, and permeability is also sensitive to cracks and fractures. It is dicult to predict permeability from other 1316

10

15

20

physical properties, such as porosity and grain size because of the complexity of the relationship. Permeability measurements under high conning pressure in a laboratory are a possible way to predict permeability structure at depths. One of the simple methods for the permeability measurement is steady state method, and using gas as a pore uid gives us the following advantages: (1) Commercial gas ow-meters that cover wide range of ow rate for any kind of gas is available so as to measure wide range of permeability quickly and accurately. (2) Nitrogen gas is chemically inert, and it is only necessary to consider the mechanical eects for the permeability changes with and increase of conning pressure. (3) Gas is less sensitive to temperature change that may induce the error for permeability measurement compared to water. Permeability does not depend on pore uid, therefore permeability measured by gas should be the same with that measure by water. Therefore several uid ow problems were analyzed using permeabilities that were measured by gas instead of water (Wibberley, 2003; Noda and Shimamoto, 2005; Wibberley and Shimamoto, 2005). However, it has been argued that (1) permeability to gas changes with pore pressure changes at same eective pressure, and (2) permeability to gas is relatively higher than that to water. These problems might be related to the pore pressure dependence on gas permeability of Klinkenberg eect (Klinkenberg, 1941). There are few examples, though, that permeabilities are compared with dierent pore uids for the same samples (Brace, et al.,1968; Faulkner and Rutter, 2000). Therefore, in this study, permeability measured by nitrogen gas as pore uid is compared with that measured by water on same samples in eective pressure cycling tests. Then Klinkenberg eect for gas permeability is discussed from experimental data. Sedimentary rocks in the western foothills of Taiwan are used for permeability measurements. 2 Intrinsic permeability and Klinkenberg eect

25

Intrinsic permeability represents a mobility of uid within porous rock materials and solely related to pore geometry of rocks itself (porosity, pore shape and pore size dis1317

tribution etc.), and is independent of uid property. Therefore, intrinsic permeability measured by any gas as pore uid in a laboratory test should be same with that measured by any kinds of uids. Permeability is functions of both medium and a uid property, and its relationship to intrinsic permeability is given by
5

K =

k g

(1)

10

15

where K is permeability (hydraulic conductivity) (m/s), k is intrinsic permeability (m2 ), is viscosity of uid (Pas), is uid density (kg/m3 ), g is gravitational acceleration 2 (m/s ). Equation (1) indicates that uid mobility depends on both uid viscosity and density. Klinkenberg (1941) discovered that permeability to gas is relatively higher than that to water, and he interpreted this phenomena as slip ow between gas molecules and solid walls. Gas molecules collide each other and to pore-walls during traveling through the pore medium. When the pore radius approaches to the mean free path of gas molecules, the frequency of collision between gas molecules and solid walls increases. Therefore this additional ux due to the gas ow at the wall surface, which is called slip ow, becomes eective to enhance the ow rate. This phenomenon is called Klinkenberg eect, and its eect is expressed as follows, kg = kl 1 + b=
c 2r 3 4cl r

= kl 1 +

c T 2r 3 p

= kl 1 +

b p

(2)

where kg is permeability to gas (m2 ), kl is permeability to liquid (m2 ), l is mean free

20

path of the gas molecules (m), r is pore radius (m), is Boltzmanns constant (JK1 ), T is temperature (K ), c is constant, p is pore pressure (Pa), b is Klinkenberg slip factor (Pa). The third term in Eq. (2) is given by the following relationship, l= T 2p 4 2r 1318 (3)

10

15

(Furbish, 1997). Equation (2) indicates that if pore radius and gas pore pressure are small and temperature of gas is high, kg becomes much larger than kl , and kg approaches to kl when pore pressure goes to innite. Brace et al. (1968) and Zoback and Byerlee (1975) performed gas permeability tests at high pore pressure conditions (p=5 to 15 MPa) by transient ow method as to avoid the Klinkenberg eect. Brace et al. (1968) showed that permeability to water on Westerly granite was similar to that to argon gas, however they did not suggest how large pore pressure is enough to avoid Klinkenberg eect for gas permeability tests. Jones (1972) investigated the relationship between Klinkenberg slip factor b and estimated water permeability kl for about 100 core samples, and he found that the log 12 17 2 b was linearly increased with the log kl at the range from 10 to 10 m . Perso and Hulen (2001) estimated the relationship between b and kl for less permeable metagraywacke (1018 to 1022 m2 ), and their results were generally consistent with Jones, though the slope seems steeper than Jones line and their data are not enough to convince the feature (8 points). 3 Geological information for specimens

20

25

Surface derived sandstones in the west-central part of the western foothills of Taiwan (Pleistocene to Miocene) were used for our permeability measurements (Table 1). The western foothills is known as a fold and thrust belt, which is developed by the Luzon arc-Asian continental collision during Miocene (Lee, 1990). 1999 Taiwan Chi-Chi earthquake was occurred at the central part of this area, and one of the main thrust faults, the Chelungpu Fault is activated during the earthquake. The western foothills is also well known as oil and natural gas elds, and overpressure is observed at depths in the several oil wells (Suppe and Wittke, 1977). Therefore permeability at depth of fault zones must act important roles for the earthquake behavior, and overpressure generation and maintenance must be related to permeability structure of the basin. Both water and gas permeabilities are also important parameters for oil and natural gas storage 1319

and migration mechanisms. 4 Experimental apparatus and permeability measurement

10

All samples were cored and polished to cylindrical shapes. All specimens have lengths of 5 to 40 mm and diameters of approximately 20 mm. Samples were dried at 80 C in the oven to eliminate pore water for gas permeability measurement, without removing structured water adsorbed to clay mineral surfaces. The experiments were performed by an intra-vessel oil pressure apparatus at room temperature. Experimental system is described in Fig. 1 and the oil apparatus can be built conning pressure up to 400 MPa by mini pressure generator. To conne pressure, avoid the leak, and eliminate the lateral ow during permeability tests, samples are jacketed by 3 to 4 layers of polyolen that can be easily contracted by heating. To evaluate intrinsic permeability k, the linear form of Darcys law is applied, and it is written as k Q = P Pdown A L up (4)

15

20

where Q is the volume of uid measured per unit time, A is the cross-sectional area of the sample, is viscosity of the pore uid, L is sample length, and Pup and Pdown are pore pressure of the upper and lower ends of the specimen, respectively. Equation (3) can be applied to incompressible media, however when compressible gas (PV = constant) is used as pore uid, the average gas permeability kgas is expressed as (Scheidegger, 1974)
2 Q kgas Pup (Pdown ) = . A L 2Pdown 2

(5)

Therefore Eq. (3) is used to calculate permeability to water and Eq. (4) for nitrogen gas in our experiment. 1320

10

15

20

Permeability was measured by the steady-state ow method that dierential pore pressure, Pup Pdown , is kept constant and that the ow rate owing out from the downstream of the samples is monitored. Upstream side of pore pressure, Pup , was controlled by the gas regulator for both gas and water permeability measurement, and uid that ows out from the downstream of a specimen was released to atmospheric pressure, which Pdown was assumed constant of 0.1 MPa. Viscosity of nitrogen gas 6 3 and distilled water was assumed constant of 16.610 (Pas) and, 1.010 (Pas), respectively. Wide range of gas ow rate from 5 to 0.024 ml/min, which corresponds to 13 2 19 2 permeability from 10 m down to 10 m in our experiments, was easily measured by coupling several gas ow meters. On the other hand, the ow rate of water was calculated using the digital balance to monitor the weight increment of water owing out from specimen continuously. Therefore there is no lower limit for water permeability measurement if time permitted. The conning pressure was cycled from 5 MPa to the maximum 160 MPa for permeability tests. To compare permeability to nitrogen gas and that to water in the same specimen, nitrogen gas was used as a pore uid for the rst 2 or 3 pressure cycles, then distilled water was used instead of nitrogen gas for the last 2 pressure cycles. Before switching pore uid from nitrogen gas to water, pore ow line was vacuumed for a while, then CO2 was lled in the pore ow line. After vacuumed again, water is poured into the specimens. As solubility of CO2 to water is quite large under high pressure, we can minimize the inuence of gas remained within pore and pore uid lines for the water permeability measurement. Pore pressure dependence on permeability were also investigated to change the dierential pore pressure at the same conning pressures. 5 Experimental result

25

The results of pressure cycling tests for sandstone specimens are shown in Fig. 2. As permeability varies signicantly with the change of pore pressure, and error-bars in 1321

10

15

20

25

gures mainly represent the dierence caused by the pore pressure variation (though pore pressure dependences were not measured in several points). The accuracy of ow rate is less than 1%, therefore errors of each plot are within the symbols. In IVA407, 5 cycles of the eective pressure test are conducted. For rst three eective pressure cycle, nitrogen gas was used as a pore uid, then distilled water was used for the other 2 pressure cycles. Permeability decreases with an increase of eective pressure in all pressure cycles, though the permeability reduces less than one order of magnitude even at 100 MPa. To compare the permeability at the same eective pressure and dierential pore pressure (IVA407 in Fig. 2 plots permeability at 0.2 MPa of Pup Pdown ), permeability to nitrogen gas is 2 to 4 times larger than that to water, and the permeability change due to pore uid dierence is largest compared to that due to change in eective pressure and its cycles at the same pore uid. The other three specimens (IVA418, IVA419, and IVA478) showed similar trend with IVA407 that permeability to gas is larger than that to water at the same dierential pore pressure. However the ratio of permeability to nitrogen gas and water diers between samples. In the case of IVA419 that permeability is relatively smaller than IVA407, permeability to gas is 5 to 7 times larger than that to water. For IVA418 and IVA478 which permeabilities are lowest, permeability to gas is more than 10 times larger than that to water. These results imply that the permeability dierence between nitrogen gas and water is larger for less permeable specimens. Figure 3 illustrates the examples of the relationship between permeability to nitrogen gas and dierential pore pressure. Generally, permeability decreases with an increase of dierential pore pressure at all eective pressures. For IVA419, the maximum permeability dierence among 0.2 to 0.8 MPa of dierential pore pressure is by the factor of 1.6, and the dierence is nearly the same with an increase of eective pressure. Figure 4 describes the examples of the relationship between permeability to water and dierential pore pressure. Permeability increases with an increase of dierential pore pressure for IVA419, which indicates opposite trend to permeability to gas. The maximum dierence in permeability due to the dierential pressure change is by the 1322

factor of 1.5. However, for IVA478, the dierential pore pressure dependence on permeability is not clear. 6 Discussion

6.1 Klinkenberg eect

5

10

15

In Fig. 5, permeability to nitrogen gas is plotted against the inverse of the average pore pressure, Pav , (Pav =(Pup +Pdown )/2) to conrm Klinkenberg eect. For most cases, permeability to gas increases linearly with an increase of 1/Pav , and this is consistent with Klinkenberg equation on the assumption that average pore pressure is equivalent to pore pressure of p on Eq. (2). Permeability to water kl can be estimated from these slopes. Figure 6 shows the dierence between estimated permeability from Klinkenberg equation and actual permeability to water. IVA478 indicates that both permeability is almost the same values. For IVA407, IVA418 and IVA419, estimated water permeability is smaller than actual ones, though the dierence becomes much smaller compared to the dierence between water and gas permeabilities, except IVA407. These results suggest that dierence of permeability between gas and water is strongly affected by Klinkenberg eect. For IVA407, estimated water permeability is disordered, and this might be caused by Turbulent ow at higher ow rate for permeability measurements. 6.2 Klinkenberg slip factor

20

Not only water permeability but Klinkenberg slip factor b can be also approximated from the tting slopes in Fig. 5. The relationship between water permeability and constant b, both of which are estimated from Eq. (2), for 30 sedimentary rocks at the western foothills (330 points) are plotted on Fig. 7. Though the points are scattered, b decreases with increasing permeability to water. The relationship in Fig. 7 is described 1323

as, b = B klC (6)

10

where B and C are constants. When units of kl and b are selected as m2 and MPa respectively, the constants B and C are evaluated as 1.50.6107 and 0.370.038, respectively. The constant C is consistent with previous study by Jones (1972) (C=0.36). The constant B becomes 7.00.4 when kl and b are expressed in the unit of md and psi, and the value is also consistent with Jones (1972) of B=6.9. However, under 19 2 10 m of permeability, the slope seems to become steeper, and this trend is consistent with Perso and Hulen (2001). If Eq. (5) is suitable to any sedimentary rocks, the relationship between the ratio of permeability to gas and water kg /kl and dierential pore pressure Pup Pdown in sedimentary rocks could be illustrated in Fig. 8. Constant B=1.5107 and C=0.37, which were approximated from Fig. 7, were applied. When permeability to water is 16 2 higher than 10 m , the ratio of permeability to gas and water is less than 2. However 17 2 below 10 m of permeability to water, the gap becomes larger, and the ratio kg /kl is 30 at 1020 m2 of kl and 0.05 MPa of dierential pore pressure. Figure 7 implies we cannot neglect the dierence between permeability to gas and water any more when permeability is small, and smaller dierential pore pressure is applied for permeability measurements.

15

20

6.3 Pore pressure dependence on water permeability As shown in Fig. 4, several specimens show dierential pore pressure dependence on permeability to water. Sasaki et al. (2003) suggests this trend is due to Bingham plastic ow within small pores. Byerlee (1990) insisted that uid that ows within the impermeable fault zone may act as Bingham ow, and this may increase the uid pressure in the fault zone, and reduce the fault strength. Buckingham-Reiner equation which is a basis of Bingham ow formula for one capillary tube model is described as (Massey 1324

25

and Smith, 1998), Q = pR 4 / 8Lp 1 4/3 p0 /p + 1/3 p0 /p

4

(7)

where R is radius of capillary tube (m), p0 is initial dierential pressure that uid ow occurs (Pa), p is dierential pore pressure (Pa), L is a length of a ow path (m), p is the slope of the relationship of shear stress against rate of shear when p >p0 (Pas). To combine Darcys law in Eq. (1) with Eq. (6), permeability to Bingham plastic uid is expressed as a following equation, kl = Qw L Ap = R 2 8 1 4/3 p0 /p + 1/3 p0 /p
4

(8)

10

15

20

25

where p =w is assumed. The experimental result in IVA 389 is compared to Eq. (7) which. all pore sizes are assumed to be same (Fig. 9). p0 and R that are evaluated by least square method became 0.047 MPa and 0.049 m, respectively. Approximation curve is generally consistent with experimental data that permeability increases with an increase of dierential pore pressure, though Eq. (7) does not completely reproduce. This might be caused by the assumption that all pore sizes in the specimen are same, which are far from natural sedimentary rocks. Furthermore, approximated pore size of 0.049 m is quite large compared to general sandstone. Therefore accurate measurements of pore size distribution is necessary to discuss further. However, several data do not show the dierential pore pressure dependence on water permeability clearly (Fig. 4), and to convince our assumption that Buckingham-Reiner equation is correlated to this positive pore pressure dependence, additional information such as porosity and grain size are also necessary. Our laboratory experiments have focused on the porous sedimentary rocks, though it is questionable whether other rocks (incohesive rocks, granite, fractured rock etc.) show the same features. Faulkner and Rutter (2000) compared the permeability to argon gas and that to water for clay rich fault rocks. Their data suggested that argon gas permeability is 10 times lager than that to water, which is in general consistent 1325

with our results. However they concluded the dierence is caused by the reduction of eective pore diameter due to the adhesion of water molecules on crystal surface rather than Klinkenberg eect or the behavior of Bingham plastic ow. It is possible that several mechanisms might aect mutually for water and gas ow, and the most dominant factor depends on the characteristics of specimens, such as the pore shape, mineral components and permeability. 7 Conclusions

10

15

20

25

Permeability was measured by nitrogen gas and water on the same specimens in eective pressure cycling tests, and followings are remarkable results in our study. (1) Permeability to nitrogen gas was 2 to 10 times larger than that to water on the same specimen. (2) Permeability to nitrogen gas decreases with an increase of pore pressure (or dierential pore pressure), and the relationship between permeability to gas and pore pressure could suit to Klinkenberg equation for most experimental data. Moreover water permeability estimate. (3) The relationship between Klinkenberg slip 14 20 2 factor b and water permeability at the wide range from 10 to 10 m is consistent 12 with the empirical relationship reported by Jones (1972) for permeable range (10 to 17 2 10 m ) that the log b decreases linearly with an increase of log water permeability. (4) Permeability to water shows positive relationship to dierential pore pressure, which can be partially explained by Buckingham Reiner equation, though the approximation curve does not fully reproduce experimental data. Our experimental results suggest that the dierence between permeability to gas and water is explained by Klinkenberg eect. Therefore if the permeability measured by compressible gas is substituted for water permeability, it is preferable to revise gas permeability using Klinkenberg equation. Moreover it is better to include pore pressure dependence on both water and gas permeabilities for uid ow problems. If we want to avoid Klinkenberg eect during steady state gas permeability measurements, large differential pore pressure above 1 MPa should be applied, though its value is not enough 1326

for impermeable sedimentary rocks. References

Brace, W. F., Walsh, J.B., and Frangos, W. T.: Permeability of granite under high pressure, J. Geophys. Res., 73, 22252236, 1968. Byerlee, J.: Friction, overpressure and fault normal compression, Geophys. Res. Lett., 17, 21092112, 1990. Faulkner, D. R. and Rutter, E. H.: Comparisons of water and argon permeability in natural claybearing fault gouge under high pressure at 20 degrees C: J. Geophys. Res., 105, 16 415 16 426, 2000. Furbish, D. J.: Fluid physics in geology; an introduction to uid motions on earths surface and within its crust, Oxford university press, New York, 1997. Jones, S. C.: A rapid accurate unsteady state Klinkenberg parameter, SPE Journal, 383397, 1972. Klinkenberg, L. J.: The permeability of Porous media to liquids and gases, American Petroleum Institute, Drilling and Productions Practices, 200213, 1941. Massey, B. and Smith, J. W.: Mechanics of uid, Spon press, London and New York, 1997. Noda, H. and Shimamoto, T.: Thermal pressurization and slip-weakening distance of a fault; an example of the Hanaore Fault, southwest Japan., Bulletin of the Seismological Society of America, 95, 12241233, 2005. Perso, P. and Hulen, J. B.: Hydrologic characterization of reservoir metagraywacke from shallow and deep levels of The Geysers vapor-dominated geothermal system, California, USA, Geothermics, 30, 169192, 2001. Sasaki, T., Watanabe, K., Lin, W., and Hosoya, S.: A study on hydraulic conductivity for Neogene sedimentary rocks under low hydraulic gradient condition, Shigen-to-Sozai, 119, 587 592, 2003. Scheidegger, A. E.: The physics of ow through porous media, 3rd Edition. University of Toronto Press, Toronto, 1974. Suppe, J. and Wittke, J. H.: Abnormal pore-uid pressures in relation to stratigraphy and structure in the active fold-and-thrust belt of northwestern Taiwan, Petroleum Geology of Taiwan, 14, 1124, 1977.

10

15

20

25

30

1327

10

Teng, L. S.: Geotectonic evolution of late Cenozoic arc-continent collision in Taiwan, Tectonophysics, 183, 57-76, 1990. Tsutsumi, A., Hirose, T., Mizoguchi, K., Uehara, S., Sato, K., and Shimamoto, T.: Measurement of Gas-permeability of rocks using an oscillating pore pressure method and a gas ow method an example of the Kakegawa Group siltstone, Shigen-to-Sozai, 119, 514518, 2003. Wibberley, C.: Hydraulic diusivity of fault gouge zones and implications for thermal pressurization during seismic slip, Earth, Planets and Space, 54, 11531171, 2002. Wibberley, C. and Shimamoto, T.: Earthquake slip weakening and asperities explained by thermal pressurization, Nature, 436, 7051, 689692, 2005. Zoback, D. M. and Byerlee, D. J.: The eect of microcrack dilatancy on the permeability of westerly granite, J. Geophys. Res., 80, 752755, 1975.

1328

Table 1. General information for specimens.

Specimen symbo Formation Burial depth (m) 39004500 45004800 45004800 39004500 Age (Ma) 11.614.5 17.819 1415 13.918 Length (mm) 20.8 12.4 23.4 18.4 Diameter (mm) 20.6 20.7 20.3 20.7 Initial porosity (%) 22.7 13.6 20.6 8.5

IVA407 IVA418 (IVA389) IVA419 IVA478

Tungkeng Fm Shihmentsum Fm Kuanyinshan Ss Hourdonqkeng Fm

1329

pressure gauge

water tank

pressure transducer

gas flow meter electronic balance

mini pressure generator porous spacer sample poly-olefin jacket CO2 N2 valve

gas regulator

gas bottle pore fluid line oil confinement line

oil pump pressure vessel

Fig. 1. Schematic diagram of gas/water permeability measurement system at high conning pressure.

1330

IVA407
N2 gas N2 gas

IVA418

permeability (m2)

permeability (m2)

10-13
cycle pore fluid Pup-Pdown

10-15
cycle 1 cycle 2 cycle 3 cycle 4 cycle pore fluid N2 gas N2 gas water water Pup-Pdown 0.4 MPa 0.4 MPa 0.4 MPa 0.4 MPa

1 cycle 2 cycle 3 cycle 4 cycle 5 cycle

N2 gas N2 gas N2 gas water water

0.2 MPa 0.2 MPa 0.2 MPa 0.2 MPa 0.2 MPa

10-16

water water
2 10-14 0 20 40 60 80 100 0 20 40 60 80 100 120 140 160

effective pressure (MPa) IVA419

4 10-14

effective pressure (MPa) IVA478

N2 gas
cycle 1 cycle 2 cycle 3 cycle 4 cycle 5 cycle
pore fluid N2 gas N2 gas N2 gas water water Pup-Pdown 1.2 MPa 1.2 MPa 1.2 MPa 1.2 MPa 1.2 MPa

N2 gas

permeability (m2)

permeability (m2)

10-16

10-14

water

cycle 1 cycle 2 cycle 3 cycle 4 cycle

pore fluid N2 gas N2 gas water water

Pup-Pdown 0.2 MPa 0.2 MPa 0.4 MPa 0.4 MPa

10-17

4 10-15 0 20 40 60 80 100 0

water
10 20 30 40 50 60

effective pressure (MPa)

effective pressure (MPa)

Fig. 2. Relationship between permeability and eective pressure in pressure cycling tests. Nitrogen gas was used as pore uid in rst 2 or 3 pressure cycles, then distilled water was used last 2 pressure cycles. Error bars mainly reect on the pore pressure dependence on permeability.

1331

2.4

IVA418
1st cycle - pressurizing

40

IVA419
1st cycle - pressurizing
Pup - Pdown

120

IVA478
2nd cycle - pressurizing
Pup - Pdown

permeability (10-15 m2)

permeability (10-18 m2)

permeability (10-15 m2)

2.2 2 1.8 1.6 1.4 1.2 1 0 20

Pup - Pdown 0.2 MPa 0.4 0.6 0.8

30

0.2 MPa 0.4 0.6 0.8

100 80

0.4 MPa 0.6 0.8 1.0 1.2

60 40 20

20

40 60 80 100 120 140 160

10

20

40

60

80

100

10

20

30

40

50

60

effective pressure (MPa)

effective pressure (MPa)

effective pressure (MPa)

Fig. 3. Dierential pore pressure dependence on permeability to nitrogen gas.

1332

0.4

IVA418
1st cycle - pressurizing

14

IVA419
3rd cycle - pressurizing

16

IVA478
4th cycle - pressurizing

permeability (10-15 m2)

permeability (10-15 m2)

permeability (10-18 m2)

0.3

Pup - Pdown 0.2 MPa 0.4 0.6 0.8

12 10 8 6 4 2

Pup - Pdown 0.2 MPa 0.4 0.6 0.8

14 12 10 8 6 4

0.2

Pup - Pdown 1.0 MPa 1.2 1.4 1.6 1.8

0.1

20 40 60 80 100 120 140 160

20

40

60

80

100

10

20

30

40

50

60

effective pressure (MPa)

effective pressure (MPa)

effective pressure (MPa)

Fig. 4. Dierential pore pressure dependence on permeability to water.

1333

IVA418
2.0
1st cycle - pressurizing

26

IVA419
1st cycle - pressurizing
60

0.12

IVA478
2nd cycle - pressurizing
effective pressure

permeability (10-15 m2)

permeability (10-15 m2)

permeability (10-15 m2)

effective pressure =

1.8 1.6 1.4 1.2 1.0

10 MPa

24 22 20 18 16 14
effective pressure

80 100

0.10

= 5 MPa

20 30 40 60 80 100

0.08

10

= 40 MPa

0.06 0.04 0.02

20 30 40 50 60

1.5

2.5

3.5

4.5

5.5

1.5

2.5

3.5

4.5

5.5

1.5

2.5

3.5

2/(Pup+Pdown) (MPa-1)

2/(Pup+Pdown) (MPa-1)

2/(Pup+Pdown) (MPa-1)

Fig. 5. Relationship between gas permeability and inverse of average pore pressure to conrm Klinkenberg eect in Eq. (2).

1334

80

IVA407 permeability (10-15 m2)

permeability cycle estimated from 1 cycle Klinkenberg equation 4 cycle measured by water 5 cycle

0.4

IVA418
cycle permeability estimated from 1st cycle Klinkenberg equation 3rd cycle measured by water 4th cycle

permeability (10-15 m2)

70 60 50 40 30 20 10 0 0 20

0.3

0.2

0.1

40

60

80

100

20

40

60

80 100 120 140 160

effective pressure (MPa)

12

effective pressure (MPa) IVA478

20

IVA419 permeability (10-18 m2)

permeability cycle estimated from 1 cycle Klinkenberg equation 3 cycle measured by water 4 cycle

permeability (10-15 m2)

permeability estimated from Klinkenberg equation

cycle
1 cycle 2 cycle 3 cycle 4 cycle 5 cycle

10 8

16 12 8 4

measured by water

6 4 2

20

40

60

80

100

10

20

30

40

50

60

effective pressure (MPa)

effective pressure (MPa)

Fig. 6. Comparison of estimated water permeability from Klinkenberg equation and experimental data.

1335

Fig. 7. Klinkenberg slip factor and estimated water permeability power law relationship from the result of gas permeability test for 30 sedimentary rocks in the Western Foothills.

1336

30 25 20 Kl m 2 10 -14 10 -16 10 -17 10 -18 10 -19 10 -20

Kg / Kl

15 10 5

0.2

0.4

0.6

0.8

1.2

1.4

Pup - Pdown (MPa)

Fig. 8. The relationship between the ratio of gas and water permeability and dierential pore pressure evaluated from the empirical relationship of Fig. 8. Constant B of 1.5107 and C of 0.37 are used to illustrate curves.

1337

Fig. 9. Dierential pore pressure dependence on water permeability. The trend is compared to the approximation curve of Eq. (7), which is based on the Bingham plastic ow into the assembly of a capillary tube.

1338